PENERAPAN AUTOREGRESSIVE DISTRIBUSI LAG (ARDL) PADA PREDIKSI PRODUKSI KAKAO INDONESIA

Abstract

Cocoa is an important commodity for Indonesia. The area and production of cocoa during the last decade decreased by 0.39% and 0.41% per year, respectively. This study aims to find a model and predict Indonesian cocoa production. This study using Autoregressive Distributed Lag (ARDL) method. The ARDL model (1, 3, 2, 2) was obtained which was selected based on the smallest Akaike Integration Criteria (AIC) value. Based on the ARDL model, the estimated average production increases by 0.6684 tons for a decrease in production of 1 ton at t-1. The estimated average production increased by 0.6130 tons for an increase in plant area of 1 Ha at time t, an increase of 0.0257 tons for an increase in plant area to produce 1 Ha at t-1, an increase of 0.5043 tons for an increase in plant area produce 1 Ha at t-2, an increase of 0.3308 tons for an increase in plant area to produce 1 Ha at t-3. Likewise for the increase or decrease area of immature plants and damaged plants in the ARDL model can be interpreted in this way.